RE: [PANIC, hyperv] BUG: unable to handle kernel paging request at ffff880077800004 (hv_ringbuffer_write)

[Dexuan Cui <decui@microsoft.com>]

> -----Original Message-----
> From: Dan Carpenter
> Sent: Friday, August 29, 2014 17:50 PM
> To: KY Srinivasan
> Cc: Sitsofe Wheeler; Dexuan Cui; Greg Kroah-Hartman; Haiyang Zhang;
> linux-kernel@vger.kernel.org; devel@linuxdriverproject.org
> Subject: Re: [PANIC, hyperv] BUG: unable to handle kernel paging request at
> ffff880077800004 (hv_ringbuffer_write)
> 
> It's sort of weird that this is an alignment issue.  The config that
> was posted earlier was an x86_64 config and on that system then the
> pointers from kmalloc() should already be aligned correctly at
> sizeof(u64).
> 
> Earlier I said I suspected the code is racy, so possibly changing from
> kmalloc() to get_zeroed_page() just changes the timing enough so we
> don't see the race condition.
> 
> regards,
> dan carpenter

Hi Dan,

Section 4.11.3 of Hypervisor Top Level Functional Specification(available at
http://blogs.msdn.com/b/virtual_pc_guy/archive/2014/02/17/updated-hypervisor-top-level-functional-specification.aspx)
says HV_STATUS_INVALID_ALIGNMENT can also be returned for a hypercall if
"the specified input or output parameter lists spans pages."

My understanding is: the input's memory range shouldn't cross a page
boundary.

Thanks,
-- Dexuan